To steer the sustainable transition in the food and energy sectors, reliable environmental data is required to answer environmental questions related to single agricultural crop, food product or energy technologies. Life cycle assessment (LCA) has been widely applied to assess the environmental footprints and mitigation potentials of a product. Given the large spatial variabilities of food and electricity production, regionalized LCA is regarded to provide more accurate environmental impact.
When the sourcing country of production origin for a purchased product is unknown, a process-based regionalized LCA is often conducted arbitrarily with subjective choices of estimating sourcing countries of production origins. This thesis developed a general process-based regionalized LCA computational structure to improve the inclusion of spatial details of tracing the spatial locations of cross-border product flows from origin of production to destination of consumption, based on the commodity balancing of a product on the country level. The model is validated with a numerical example and demonstrated with a case study from literature for an improved accuracy of impact results. The proposed model offers a coherent and transparent way of analyzing the influence of different trade assumptions or truncation errors. It can be used to improve the global value chain modeling of agricultural commodities.
Increasingly, companies are making product footprint and comparative claims available on the individual product level. International food companies often have a global footprint in their product supply chain and a large product portfolio for the same functionality sold in various consumer markets. This thesis developed a stepwise framework for operationalizing the application of regionalized LCA to assess a large-scale portfolio of food product. Its feasibility and reliability are tested with a case study comparing 212 plant-based fat spreads and 40 dairy butters sold in 21 countries. It shows large inter-product variabilities, ranging from 0.98 to 6.93 (mean 3.3) kg CO2-eq/kg for 212 plant-based spreads and 8.08 to 16.93 (mean 12.1) kg CO2-eq for 21 dairy butters. The key drivers and main uncertainties of impact are the assumptions of the sourcing country of production and GHG emission from land use change.
This thesis further assessed the influence of different regionalized LCA model assumptions and temporal resolutions on the carbon footprint of power to gas (PtG) applications. When the electricity input is based on a renewable electricity mix with guarantee of origin, PtG under study have a 32-65% reduction of carbon footprint compared to fossil natural gas. With current national average consumption mix on a yearly basis, PtG production in Switzerland could be operated to provide climate benefits. However, when moving from yearly average to hourly resolution, PtG has a higher carbon footprint for more than 50% of time over the year. Thus, the deployment of PtG should be guided in a finer temporal resolution to gain potential climate benefits.
The regionalized LCA model and methodology as well as case studies contribute to advance our understanding in the methodological aspect of regionalized LCA model and key issues related to its practical operationalization and applications.
EPFL_TH9688.pdf
n/a
openaccess
Copyright
7.15 MB
Adobe PDF
d47542530eb12b4ad56d732faa19665f